摘要
针对骨髓细胞图像的特点,采用连续小波变换对图像进行了处理,在消除原始图像噪声的同时,从不同的角度检测出图像的主要边缘。采用两级神经网络,利用基于神经网络的GHA算法获得图象的三个主分量,然后采用模拟退火算法和BP算法进行细胞的分类识别,获得了较好的识别效果。
Aimlng at the characteristics of the marrow cell image, the paper adopts continuous wavelet transform to process image so that main edges can be detected from different angle in the meanwhile original image noise being eliminated. Adopt two neural networks. Using Principal Component Analysis based neural network(GHA) to acquire three principal components and using Simulated Annealing(SA) and BP network to class and recognize the marrow cells.
出处
《计算技术与自动化》
2005年第3期57-59,共3页
Computing Technology and Automation
基金
内蒙古高等学校科学研究项目(NJ04109)